LabAnalyst has six menus, plus on-line instructions in the "Help" menu (OS 8 and above) or the Apple menu (earlier operating systems).
| FILE | EDIT | ANALYZE | VIEW | CHANNELS | SCRIPTS | HELP |
GENERAL NOTES ON ANALYSIS PROCEDURES:
DATA BLOCKS:
If no data block is selected, only the RUNNING
AVERAGE, PAIRS DIFFERENCES, and FFT analyses are available (and
except for FFT, only in Active Screen mode). For all other
analyses a data block must be selected.
To select a data block:
- Use the normal Macintosh click-hold-and-drag method, with the cursor within the plot area.
- Move the cursor to the beginning point on the plot area and click once, then repeat this procedure for the end point (which may be on either side of the beginning point).
- Alternately, blocks of predefined duration can be selected with single clicks (BLOCK WIDTH option, below).
Blocks can include multiple screens. In ACTIVE SCREEN mode, mark the start of the block, then shift screens as necessary until the end of the block can be marked.
Once you have performed
an analysis operation and want to select a new data block, click the 'Exit'
or 'Cancel' button in the RESULTS window (or that window's
close box), or click on any other window. The cursor will become a
crosshair when within the plot area.
This example of a block
window shows the upper and lower limits of the data within the block.
This particular block (being analyzed with the Integrate option)
also shows:
MORE BLOCK HANDLING AND ANALYSIS TOOLS:
When activated,
the AUTOREPEAT option will repeat the most recently used analysis
whenever you select a new data block. This simplifies doing the same
analysis on a series of different blocks, but it can be annoying if you
wish to perform different analyses. If autorepeat is on and you want
to do something else, pick the new operation from the ANALYSIS menu
or from the 'toolbar' buttons at the bottom of the plot area.
For most ANALYSIS
operations, clicking the plot area is equivalent to clicking a highlighted
button.
Scaling factors
may be applied to results for several ANALYSIS operations with the
'scale' button (see below).
The 'Store'
button in many of the analysis mode windows allows you to directly transfer
the current results mean for use as a scaling factor. When you click
the 'Store' button, the scaling factors window appears. Click
on any channel's ' * ' or ' ÷ ' button, and the current
mean will appear.
If no data block is selected, you can perform the following two operations (as well as Fast Fourier transforms):
RUNNING AVERAGE... Lets
you click on a series of SINGLE points while LabAnalyst keeps track
of the cumulative mean, standard deviation (SD), coefficient of variation
(CV), and standard error (SE). The program also shows the summed total
of all selected points, and the range (minimum and maximum values).
To select points for analysis, move the cursor to the point of interest in the plot area and click once. Then move to the next point and repeat.
PAIRS DIFFERENCE...
Calculates the cumulative mean, SD, and SE of a
series of PAIRS of points. You select point pairs with mouse clicks
(an example might be the peaks and valleys of a ventilation trace).
LabAnalyst measures the absolute difference between the points in
the pair. It also calculates the cumulative mean, SD, and SE of the
time intervals between points. As for RUNNING AVERAGE, the
delete key gets rid of the last pair of points selected.
BASIC STATS 
+A 
Computes the mean,
range, standard deviation (SD), standard error (SE), and coefficient of
variation (CV) of the block (it also shows the time duration and the data
range).
The 'distribution'
button produces a bar graph of frequency distribution. If the file
has more than one channel, you can click on buttons for different channels
and get those means (you can also use the keyboard to select channels).
To get the integrated total for a rate function, such as oxygen consumption, make sure the correct rate unit is selected (i.e., 'per min' if the units are ml/min or 'per hour' if the units are ml/hour). A more sophisticated integration mode is available in the INTEGRATE BLOCK option.
Use the 'limits' buttons to restrict analysis to a subsection of the block (move the cursor to the block window and click when the limit line is correctly positioned).
The 'scale' button toggles scaling of the results (see SCALE RESULTS, below). The 'Store' button in this and other analysis mode windows allows you to directly transfer the current mean for use as a scaling factor. When you click 'Store', the scaling factors window appears. Click on any channel's "*" or "÷" button, and the current mean will appear in the first edit field (the multiplication or division factor) for that channel.
MINIMUM
VALUE... +M
MAXIMUM
VALUE... 
MOST
LEVEL... MOST
VARIABLE...
Allow you to enter an interval (in seconds) and LabAnalyst will search within the selected block for either the highest or lowest continuous average over that interval, or the most level or most variable region within the block.
Activating the 'reuse
this interval' button in the interval selection window will bypass the
interval selection routine during subsequent uses of these analyses (such
as when using the AUTOREPEAT option for new blocks).
For the MOST LEVEL option, the 'most level' region is the interval in which the sum of absolute differences from the interval mean (i.e., the sum of Xi - meanX, or point-to-point variance) is lowest. Note that this is not necessarily the interval with the lowest slope, although this usually turns out to be the case.
For the MOST VARIABLE option, you can select to search either for the region with maximum overall slope or the region with maximum point-to-point variance.
When your interval selection is complete, click the 'interval OK'
button and the program will find the appropriate interval and display
the results in the next window, shown below.
After calculations, the
maximal, minimal, or most level area is shown as a color-inverted rectangle
on the block window.
As for BASIC STATS, you can switch to other channels, but in the default mode the interval boundaries remain constant -- i.e., the same beginning and ending points as on the initially scanned channel are used for other channels. This sounds confusing but it allows you to scan for the period of, say, lowest VO2 and then get the temperature, CO2, etc. for that specific period.
Alternately, you can activate the 'rescan new channels' button to force a re-scan of each new channel selected.
Two other considerations:
When using
the MINIMUM VALUE... and MAXIMUM VALUE... functions,
a button labeled 'C.V.R. options...' is available. This
stands for Constant Volume Respirometry, and it opens
a window that lets you set up the variables needed to compute O2
or CO2 exchange in a closed system.
Note that this option assumes that the data being analyzed are in units
of % gas concentration and that baseline has already been corrected.
You need to specify the gas type, the chamber volume, the elapsed time, the chamber temperature, the barometric pressure, the initial relative humidity in the chamber (if the gas contains water vapor), the initial concentrations of O2 and CO2 (FiO2 and FiCO2), and the respiratory exchange ratio (RQ). You also need to specify whether or not CO2 is absorbed prior to oxygen analysis ('excurrent CO2' buttons). When done, click the 'Selection OK' button.
When C.V.R. options is activated, the results window shows gas exchange rates in units of ml/min -- but note that only the mean value is computed as gas exchange (the SD, SE, etc. are shown in their original units). To switch off the C.V.R. calculations, click the 'C.V.R. options...' button.
INTEGRATE BLOCK...
Finds
the area under the curve within the defined block.
There is a choice of
time units (seconds, minutes, hours) and baseline values. The baseline
for integration can be set at zero, the initial value of the
block, the final value of the block, a user-specified value,
or a proportional linear correction between initial and final block
values. Because of the different baseline options, this operation
is considerably more versatile than the integration feature including in
BASIC STATS, MINIMUM, MAXIMUM, and LEVEL.
In this example, the block is integrated using the zero baseline option with the time unit set as minutes. Note that the start, end, and proportional options apply to the block defined with the left and right limit functions.
Keep in mind that you need to pick the time unit that matches the rate unit used in the channel being analyzed, or the results will not be valid. If you are using a rate unit for which a matching time unit is not available, you will need to adjust the results manually. For example, if your data are in units of Kilojoules/day, you will need to divide the results by the factorial difference between days and whatever unit you select. To continue with this example, if you set the units to 'hours' and your data are in KJ/day, you must divide the results by 24 (since there are 24 hours per day). Similarly, if you set the units to 'minutes', you would need to divide by the number of minutes per day (1440). This can be done conveniently using the 'Store' and 'Scale' buttons.
If you select the "Integration
plot" option (button in the bottom row), the computer generates
a plot of the integrated values over time. This plot will change whenever
a new channel, right or left limit, or baseline option is selected.
Additional considerations:
WAVEFORM...
+W
Uses
simple algorithms to calculate wave frequency and mean peak height.
Basically, the program looks for successive 'peaks' and 'valleys'
in the data. A peak (the 'crest' of a waveform) is defined as a series
of 5 points with the middle point having the highest value, and with values
declining in the two points on either side. The inverse is true for
a valley (the 'trough' of a waveform). LabAnalyst marks the
peaks and valleys it finds with dotted (valleys) or dashed (peaks) lines
in the block window.
In the default mode the program
will use all the data within the block. Alternately, 'filtering' is
possible through cursor selection of minimum peak and valley values ('triggers')
in the block window. Click the 'use cursor' button and move
the cursor to the block window. A horizontal line will track the cursor's
movement and a readout in the Results window will show the actual height
of the cursor. Click once to select a peak (done first) or
valley trigger. After selection, peak triggers are shown as pink lines,
and valley triggers are green lines. The post-peak trigger
value (default zero) is the number of cases the program 'skips' after finding
a peak. This option can be useful when analyzing noisy files.
A typical results window is shown at right. Note that in this example the number of peaks and amplitudes is correctly matched. The program beeps and prints a warning if no periodicity is found, if no amplitudes are found, or if the number of amplitudes does not equal the number of cycles +1 (indicating that some peaks were not associated with definable valleys, so the frequency or amplitude may be incorrect). The mean values for both peaks and valleys are also shown.
NOTE: The waveform algorithms are easily confused by noise (because of the way peaks and valleys are defined). If you are only interested in frequency, it is reasonably safe to reduce noise by smoothing data prior to analysis. However, smoothing reduces peak amplitudes (in some cases very dramatically), so it must be used with caution if you need peak height data. Smoothing is least damaging if peaks are 'rounded' and contain many more points than the smoothing interval. If necessary, use PAIRS DIFFERENCE to obtain peak heights, then obtain frequency data after smoothing.
The 'wave shape data'
button opens a window with statistical information on the waveform's rise
and decay times.
Rise time (the elapsed time from a valley to a subsequent peak) is shown in blue; decay time (the elapsed time from a peak to a subsequent valley) is shown in red. The small triangles indicate the means while the bars show the distributions.
If the data contain more than 10 peaks, additional analyses are available from the Peak and interval histograms button, which produces histograms of peak height (shown as the absolute value), wave amplitude (valley to peak) and interpeak interval (essentially the wavelength calculated on a peak-to-peak basis). An example is shown below.
The save data button stores a text file of the histogram values, the print graph button sends the data to a printer, and the square-root Y button shows the count as a square-root, which better shows bars with low counts.
| In the FP and X versions, mode selection and bar number selection are done from pop-up menus instead of buttons. |
TIME SERIES...
This option displays time series data in graphical
form. Time series analysis examines data for many kinds of temporal
relationships -- basically, does sample value at a particular time (T)
predict sample value at some later time (T + z)? A graphical display
is useful for determining if there are periodicities within the data, and
(if periodicities exist) if the waveforms are symmetrical. When this
option is selected, LabAnalyst opens a window containing edit fields
for the Start Lag (the initial time increment between samples) and
the Time Step (the time increment added for each successive plot).
Defaults are a start lag of 0 and a time step of 1 sample. Edit as
desired. There are two plotting options.
Use the 'Plot time series' button to generate 18 scatterplots. For each, the sample value (X - coordinate) is plotted against the sample value at a fixed time increment in the future (Y -coordinate):
In each plot the time increment is equal to the start lag plus the sum of the cumulative time steps (i.e., start lag + time step X plot number). The increment value is shown at the top of each plot. If there is no temporal predictability, the point distribution will be random. However, if temporal patterns exist, the distribution of points will be non-random, and the shape of the distribution will indicate the degree of symmetry and the scatter will indicate the degree of randomness or 'noise'. You can replot with different start lags and time step increments. Click the 'Print plots' button for hard-copy output. Note that 24, not 18, time steps are plotted, and that you cannot select this button until data have been plotted on the screen.
Use one of the 'step
test' buttons to generate a summarized 'periodicity test' for
50, 100, or 200 stepped time intervals (NOTE: these buttons are only available
if there are sufficient points within the block -- assuming a start lag
of 1 and a time step of 1).
This example shows a 100-step plot. Correlations with negative slopes are plotted below the zero line. Bar heights are a relative index of how value at time (T) predicts value at time (T + time step). The tallest bar (red on color screens) is the interval with highest predictability.
Click the 'Print' button for hard-copy output.
Click the 'Quit' button, another window, or the close box to return to the analysis page.
FFT... Uses a Fast
Fourier Transform (FFT) to calculate the frequency structure
of cyclic data. Theoretically, any waveform can be decomposed into
a set of simple sine waves of different frequencies and phases. The
FFT algorithm finds these fundamental frequencies and displays
them as peaks in the plot area.
Complex 'summed' frequency
data occur frequently in biology (and in other areas of science).
For example, you might want to use an impedance converter to measure the
heart rate in a small mammal, bird, or lizard. Unfortunately, in addition
to heart rate, you will also pick up signals produced by breathing movements.
Therefore the instrument output will contain a confusing summation of the
combined effects of breathing and heart rate. It may also contain
'noise' from random or irregular events (such as muscle movement from minor
postural adjustments).
The messy-looking data shown at right are an example of such a waveform. Although it is obviously complex, a visual inspection suggests that it does contain some regularity. However, this periodicity is not readily studied with either the WAVEFORM or TIME SERIES operations. Fortunately, the FFT procedure can help find the important underlying components of this complex wave. In many cases it can detect basic cycles in a data set even if they are visually 'buried' by random noise.
The first
step in the FFT procedure is selection of the size of the block of data
to be analyzed. The FFT algorithm requires that the block size be
a power of two; depending on the size of your data set you can select
any block size from 128 samples up to a maximum of 16,384 samples (in this
example, the total number of samples in the file was about 7000; accordingly,
the largest possible block for FFT analysis is 4096 samples).
After you select a block size, the program will show the block duration and then prompt you to go to the plot window and select the block to be analyzed. Do this by moving the cursor into the plot area, where it will outline a block of the size you selected. Fit the cursor block over the subset of data you wish to analyze and click the mouse once. This will select the desired FFT block.
Once the block is chosen, you can proceed transform it (Do FFT button), select another block size (by clicking the appropriate 'power of two' button), or exit. Prior to computing the FFT, you can select to show a square-root conversion of the results (this shows small peaks more clearly). You may also choose between showing a line or histogram plot of the results.
After completing the FFT,
the waveform's fundamental frequencies are shown graphically in the plot
area. You can examine the details of this structure by moving the
cursor over the plot; the fundamental frequencies that have been 'decomposed'
from the original signal, and their amplitudes, are shown numerically as
peaks in the results window.
In this example, the waveform from the first image (above) is seen to be composed of three discrete fundamental frequencies, which appear as the three sharp peaks in the plot area. The cursor is over one of the peaks, which has a frequency of 3.979 Hz and a magnitude (useful for comparisons among peaks) of .7476 (these data are displayed in the results window). Note that in current FP versions, you have a choice of output units (frequency in Hz, kHz, etc.; period in sec, min, etc.)
After the transform is complete, you can expand or shrink the display, or smooth (or unsmooth) the data (the results in this example are smoothed).
FFT results are stored in channel zero (not normally used by LabAnalyst); use the copy button to move them to a 'regular' data channel if you want to save them to disk or print them (copying is only possible if the number of 'regular' channels is <16).
Note that if you click the exit button, you are transferred to the plot area window in channel zero (which contains the FFT results). If you click the close box you are transferred back to the original data channel. If you chose the former, you can switch back to the regular channels by pushing the appropriate number key, or clicking the channel selection buttons in the upper right corner of the plot area. You cannot get back to the FFT results in channel zero except by re-running the FFT procedure.
Note that if you start the FFT procedure while using the multi-channel display mode, you'll be switched to single-channel mode (using the current active channel) prior to the beginning of analyses. At the conclusion of the FFT calculations you'll be returned to multi-channel mode IF you click the close box or (in FP versions) the plot area.
SLOPE vs TIME...
Calculates rates of change over time within a data block. Seconds,
minutes, hours, or days may be selected as the time units. You may
use a linear (least-squares) regression or a semi-log regression (log Y
= a + b*time).
When the calculations
are complete, a regression line is superimposed on the block window.
The program computes slope, r-squared, and probabilities that the slope
is either one or zero.
Two values of 'a' (the intercept) are given. One is based on the time change calculated from the start of the file (t=zero), and the other is based on the time change within the block, using the assumption that time zero is the start of the block.
You can switch channels (for new calculations) with the usual selection buttons on the bottom of the results window. To change the time units, re-select the original channel.
You can also test the slope against any user-defined value with the 'test' button.
Some additional considerations:
REGRESS CHANNELS...
This
option performs basic regression analysis for any two channels (obviously,
it is only available if the file has more than one channel):
The initial step in the regression procedure is to chose the type of unit conversions. Linear (least-squares), semi-log (log Y = a+b*X or Y = a+b*log X), and log-log (log Y = a + b*log X) regression models are available. However, some of the conversions will not be available if the data range includes zero or negative values (since one can't take the log of a negative number).
Next, you select the two channels to regress, using the window shown at right. The program won't let you attempt to regress a channel against itself, so you may have to do some fancy button-clicking to get your channels selected.
After the 'selections OK' button is clicked, LabAnalyst performs the calculations and produces a scatterplot of the data points in the block window (x values versus y values), along with the regression line.
The numerical results are
the same as for the SLOPE vs TIME option described above -- except
that only a single value of the intercept is shown.
You can test the slope against any user-defined value. Enter the slope into the edit field and click the 'test' button (very low probability values are shown as "<.00001").
Some additional considerations:
FIND ASYMPTOTE...
This is an application of first-order kinetics, in which data approach a
stable plateau (asymptote) according to a rate constant (this means
that the fractional rate of approach to the asymptote is constant
over time). Some examples include heating and cooling (the "Newtonian"
model) and gas mixing and washout characteristics (and many other physical
phenomena).
The program uses an iterative
method to find the best-fit asymptote for a selected block of data, according
to the simple model: Y = ln(asymptote - data). You can chose
any number of iterations between 6 and 50. Using a lot of iterations
might increase the accuracy of the estimate (this doesn't always
occur), but will also increase the analysis time. In practice, you
usually don't need to use more than 6 to 10 iterations for good accuracy.
Note that if you give it 'messy' data that do not conform reasonably well
to first-order kinetics, the program may take a long time to produce an
estimate (and that estimate may have fairly glaring errors).
After completing the analysis, LabAnalyst shows the asymptote, the coefficient of determination or C.D. (an estimate of the precision of the fit of the data to the model, and hence the precision of the estimated asymptote), the slope of the ln-transformed data, the rate constant (the fraction of the change between a starting value and the asymptote that is completed during 1 time unit), and the time to complete a fraction of the total change between a starting value and the asymptote (values from 1% to 99% are selectable from a pop-up menu). You can use your choice of time units (seconds, minutes, hours, or days) for slopes and rate constants.
LabAnalyst also
draws a goodness-of-fit plot that illustrates how closely the model matches
the data. Points are plotted in yellow as the log (base e) of the absolute
difference between the model predictions and the data:
Y value = ln(abs(asymptote - data))
Individual points are shown only if the total number of points in the plot is less than 60. The line predicted by the model is shown in white. In this example (not the same as for the previous figure), the analysis contained 31 points ranging in value from about 31.1 to slightly less than 0, with a predicted asymptote of -0.05.
In the latest versions, you can also plot the residuals from this regression. |
Some additional considerations:
FIT POLYNOMIAL...
(LabAnalyst FP and X only) In some cases,
relationships between different variables are best expressed with a polynomial
regression. LabAnalyst lets you fit pairs of channels with
polynomials of up to 9 degrees. The model is: Channel Y = (polynomial
expression of Channel X). After you select the X and Y
channels (as described above for linear
regression), the program computes and displays a default polynomial of 3
degrees. You can then select other degrees using the yellow pop-up
menu. In this example, a 4-degree polynomial was used to predict CO2 production from temperature; the resulting
equation -- shown at the right of the window -- explains 89% of the variance
in VCO2.
After completing the analysis, several options are available. The 'select new channels' button lets you change the X and Y variables. The 'make new chan from polynomial' button (only available if the number of channels is less than 16) generates a new channel by applying the current polynomial equation to the data in the X channel -- in other words, it computes and saves a new Y value for every sample in the X channel.
The 'show all r^2'
button helps you select the most appropriate degree for the polynomial equation.
The predictive value of a polynomial always increases as the degree
increases. However, the increase in accuracy usually plateaus, so
it is reasonable to use the simplest equation consistent with good predictive
power. When the 'show all r^2' button is clicked, the program
computes the r2 value for all degrees between 1 and 9, and then
shows a bar graph of the results. [You can remove this plot by clicking
the highlighted 'show all r^2' button.]
In the example shown at right, there was little change in predictive power after the degree of the polynomial exceeded 3 (you can select the color of bar graph plots in the Colors and lines option in the View menu.)
Some additional considerations:
TIME
INTEGRATION This operation calculates the cumulative
duration of data in the selected block that satisfy two Boolean criteria:
greater than one user-selected value, and less than a second user-selected
value.
An example time integration
window is shown at right. The default minimum and maximum values are
the lower and upper limits (respectively) of the data range in the block,
which includes 100% of the data. You can set new limits (as shown
here) in three ways:
When the 'minimum' or 'maximum' buttons are clicked, the limits are reset to these values
When the 'greater than' or 'less than' buttons are clicked, you can use the mouse to move cursor in the block window to graphically select the upper and lower limits.
Finally, you can directly type in the limit values in the edit fields (hitting "return" will force the program to recalculate the results).
Be sure to select the time unit you want. Some additional considerations include:
EVENT COUNTING
This operation counts the number of 'events'
in the selected block that satisfy two Boolean criteria: greater than
one user-selected value, and less than a second user-selected value.
An 'event' occurs when the data cross one of the two Boolean 'boundaries'.
A 'negative' event is counted when the data drop below the lower
limit, and a 'positive' event is counted when the data rise above
the upper limit.
An example of event
counting is shown at right. The default minimum and maximum values
are the lower and upper limits (respectively) of the data range in the block,
which includes 100% of the data and results in a single event in each category.
You can set new limits (as shown here) in three ways:
When the 'minimum' or 'maximum' buttons are clicked, the limits are reset to these values
When the 'greater than' or 'less than' buttons are clicked, you can use the mouse to move cursor in the block window to graphically select the upper and lower limits.
Finally, you can directly type in the limit values in the edit fields (hitting "return" will force the program to recalculate the results).
Some additional considerations include:
SELECTIVE INTEGRATION
(FP and X versions only) This operation
integrates all data in the selected block that satisfy two Boolean
criteria: greater than one user-selected value, and less than a second user-selected
value. The window shows the integrated value in absolute terms and
as the fraction of the total integrated block.
An example of selective
integration is shown at right. The default minimum and maximum values
are the lower and upper limits (respectively) of the data range in the block,
which includes 100% of the data and results in a single event in each category.
You can set new limits (as shown here) in three ways:
When the 'minimum' or 'maximum' buttons are clicked, the limits are reset to these values
When the 'greater than' or 'less than' buttons are clicked, you can use the mouse to move cursor in the block window to graphically select the upper and lower limits.
Finally, you can directly type in the limit values in the edit fields (hitting "return" will force the program to recalculate the results).
Some additional considerations include:
Analysis utilities 
In the 'FP'and 'X'
versions, this selection has a submenu with three items:
Allows automatic selection of blocks of user-defined duration with a single mouse click, instead of the normal method of clicking on both the start and end of the desired block. In this mode, the cursor is contained within a box that 'frames' the block duration in the plot area. Position the cursor over the desired block and click once to select it.
The default automatic block size is equivalent to 13 samples or 12 sample intervals (e.g., 60 seconds with a 5-second sample interval). Any other block size can be selected, as long as it contains more than 2 samples and less than 1/2 of the file duration.
Return to the standard two-click selection method with the 'Normal block selection' button.
final value = (result x B) +A or final value = (result ÷ B) +A
Different A and B values can be applied to each channel, but you must enter the correct values in any channel you wish to scale (in other words, if you have scaling factors entered for channel 5, they apply only to channel 5 unless they are also entered in another channel of interest).
This example shows:
go back to the LabAnalyst page